[Feat] 310p support MoE W8A8 quantizaition (#6641)

### What this PR does / why we need it?
This PR introduces support for W8A8 dynamic quantization for
Mixture-of-Experts (MoE) models on Ascend 310P devices. This is achieved
by:
- Implementing a new quantization scheme
`AscendW8A8DynamicFusedMoEMethod310`.
- Adding a unified MLP implementation (`unified_apply_mlp`) for 310P
that handles both quantized and unquantized paths.
- Refactoring the MoE and quantization configuration logic to correctly
route to the new 310P-specific implementations.
- Adding new e2e and unit tests to verify the functionality of MoE W8A8
quantization.

### Does this PR introduce _any_ user-facing change?
No

### How was this patch tested?
- Added a new e2e test `test_qwen3_moe_tp2_w8a8` to test MoE W8A8
quantization in a multi-card setup.
- Added several new unit tests for the 310P-specific MoE components,
including `experts_selector`, `fused_moe`, `moe_comm_method`, `moe_mlp`,
and the new `w8a8_dynamic` quantization method.

- vLLM version: v0.15.0
- vLLM main:
d7e17aaacd

---------

Signed-off-by: pu-zhe <zpuaa@outlook.com>

vllm_ascend/_310p/fused_moe/token_dispatcher.py

@@ -32,21 +32,14 @@ class TokenDispatcherWithAllGather310(TokenDispatcherWithAllGather):
     def __init__(self, **kwargs):
         super().__init__(**kwargs)
 
-    def token_dispatch(
+    def token_dispatch(  # type: ignore[override]
         self,
         hidden_states: torch.Tensor,
         topk_weights: torch.Tensor,
         topk_ids: torch.Tensor,
         expert_map: torch.Tensor | None = None,
-        global_redundant_expert_num: int = 0,
-        mc2_mask: torch.Tensor | None = None,
         apply_router_weight_on_input: bool = False,
-        with_quant: bool = False,
-        dynamic_eplb: bool = False,
-        pertoken_scale: torch.Tensor | None = None,
     ):
-        if with_quant:
-            raise RuntimeError("Quant is not supported for 310P currently.")
         self.original_shape = hidden_states.shape
 
         num_tokens = hidden_states.shape[:-1].numel()
@@ -77,7 +70,6 @@ class TokenDispatcherWithAllGather310(TokenDispatcherWithAllGather):
 
         return TokenDispatchResult(
             hidden_states=sorted_hidden_states,
-            dynamic_scale=None,
             group_list=expert_tokens,
             group_list_type=group_list_type,
             context_metadata=context_metadata,